A method has been developed by which the susceptibility of chickens to Rous virus can be tested, and the virulence of eight or ten different fluids compared in a single animal. The results of five series of experiments made with this technique can be summarized as follows: When a medium composed chiefly of chicken serum and Tyrode solution and containing no fresh tissues is inoculated with filtered extract of Rous and other sarcomas and incubated for 48 hours, it never produces a tumor after being injected into chickens. The virus has apparently been destroyed or at least has lost activity. But, in a solid medium, composed chiefly of serum and Tyrode solution and containing fragments of fresh tissues, the virus is found to increase readily, as shown in the first series of experiments. In the course of 15 months, the experiments have been repeated many times, with identical results. Flasks containing embryo pulp or leucocytes inoculated with filtered extract of sarcoma are used to keep on hand a constant supply of the Rous virus. The cultures of monocytes inoculated with the filtered extract often assume the appearance of the cultures of Rous sarcoma. They may also remain normal to all appearance despite the circumstance that the virus is multiplying within the medium. It is not certain that the activity of Rous virus is always accompanied by cell lesions, but there is no doubt that its increase depends on the presence of fresh tissues within the medium. The disappearance of the Rous virus from a medium that does not contain any fresh tissue may be interpreted as follows: the agent has been destroyed; or it is still present in a concentration lower than 1 in 50,000, which is the concentration required to produce a tumor even in the more susceptible chickens; or, according to the hypothesis of Gye, it is present in an inactive form. In the above experiments, the fresh tissues added to the medium might conceivably have enabled the virus to keep its full activity, through supplying the conditions requisite therefor, or they might merely have furnished an activating substance. The value of Gye's hypothesis was tested in a series of experiments. The results indicate that the tumor-producing virus present in the cultures was not composed of two parts, an inactive part multiplying in the medium, and an activating part supplied by the tissues. In another series of experiments, the relations between the reproduction of the virus and the quantity of the tissues contained in the medium were studied. The presence of a small fragment of leucocytic film or spleen tissue was sufficient to prevent the virus from disappearing. Approximately 1 c.mm. of spleen tissue in 3,000 c.mm. of medium may on occasion maintain a concentration of Rous virus in this fluid sufficient to produce a tumor upon inoculation into chickens. But this rarely happens. Generally when the medium contained only one fragment of spleen or leucocytic film, or 1 drop of embryonic pulp, the virus disappeared rapidly. When the quantity of tissue was from five to nine times larger, an abundant production of virus was practically always found. It became obvious that the quantity of active virus present in a medium containing multiplying cells depends upon the amount of tissue in the medium. In the fourth series of experiments, the kind of cells needed for the multiplication of the virus was ascertained. Rous virus was found to disappear rapidly from the fluid of cultures of fibroblasts, while it multiplied readily in cultures of leucocytes, the total volume of both tissues being approximately the same. It should be remembered that strains of fibroblasts obtained from Rous and other sarcomas very rarely produce tumors upon inoculation into chickens, while the inoculation of cultures of macrophages from the same tumors practically always determines their appearance. The fifth series of experiments showed that the cell metabolism is an important factor in the reproduction of the virus. When the activity of tissues had been suppressed or very much decreased by freezing, no virus was produced, while it multiplied readily in the control. The lack of oxygen for a period of 24 or 48 hours stopped cell proliferation, and at the same time the production of Rous virus ceased. However, the fluid of some of these dead or inactive cultures, after 6 days incubation, was still able to give rise to a small tumor upon inoculation into a chicken. There is an evident relation between the proliferating activity of the tissues and the production of the virus. But the agent may persist for several days in association with dead tissues. It may be concluded that the reproduction in vitro of the active virus depends on the presence of fresh tissues in the culture and upon the quantity, the activity, and the nature of the cells contained in the medium.